The inspiration for our product came from the wish to be able to help people during this challenging times, and so, we started researching and talking to people. Doing so, we’ve noticed that even the simplest aspects of their life have been changed, things as simple as having a healthy meal and the impact they have in our lives, both physically and mentally.Our app not only helps them decide and reduce food waste but It also enables them to find the recipes that better suit their tastes, and restrictions like allergies or some specific diets as well as generating awareness around their habits.
Before the ICC, we had the opportunity of building the iOS app using Swift. As the ICC started, first, we used a bit of image processing and, with the help of Tesseract OCR, we managed to recognise the text from images of invoices, and then, with a model trained in CreateML, its turned possible to link the text of the invoice to its respective ingredients.As we never had contact with machine learning tecnology, we had many challenges wiht simple things, like choosing what was the better framework to use, or even implementing the algorithms.We also bumped into some more specific problems, like making a algorithm that understands abbreviations and mistypes, and brands as well.
Some accomplishments we made during the ICC are successfully developing a beginning of a machine learning algorithm that solves the problem we had with the invoice recognition and building a more concrete idea of what we want and how we want the future of the HomeMade app.
We learned not only technical aspects, like how to develop a machine learning algorithm and solve the related problems, but also how to pitch our idea and develop it in a market sense to create something more desirable.
First is important for us to test it with real users, even without all of the possible features, because we believe that this is a very important step to know if we are in the right way. Then, we need to register more recipes and work in our recommendations algorithm, so we can provide choices specific to each user. And last but not least, make partnerships so the product is viable and we can continue developing it and make it available in other platforms.